Fig 2: mTORC2 restrains NK cell effector functions by inhibiting mTORC1 activity. a, b The expression of perforin and granzyme B (a), CD71, and uptake of 2-NBDG (b) in total splenic NK cells (CD3-CD19-NK1.1+NKp46+) and NK cell subsets from mice of the indicated genotypes. c Flow cytometric analysis of CD107a+ cells among total splenic NK cells (CD3-CD19-NK1.1+NKp46+) or the NK cell subsets specified from mice of the indicated genotypes after coculture with YAC-1 targets. d The expression of phosphorylated (p-) S6 at ser235/236 (p-S6ser235/236) in splenic NK cells (CD3-CD19-NK1.1+) from control versus Rictor?/? mice after stimulation with rapamycin of the indicated concentrations for 6 h. The MFI of p-S6 ser235/236 at each concentration was normalized to the baseline reading (i.e., 0 nm rapamycin-treated control NK cells in a littermate pair). The horizontal dashed line indicates the normalized baseline reading. e–f The expression of perforin, granzyme B, uptake of 2-NBDG (e), and CD71 and CD98 (f) (MFI relative to the littermates of the control + DMSO group) in splenic NK cells from control versus Rictor?/? mice treated with DMSO (set as the solvent control) or 0.1 nM rapamycin for 6 h ex vivo. g Flow cytometric analysis of CD107a+ splenic NK cells from control versus Rictor?/? mice following treatment with DMSO or 0.1 nM rapamycin, described in detail in the ONLINE METHODS. The control represents Rictorfl/+/Ncr1-Cre+ mice; Rictor?NK represents Rictorfl/fl/Ncr1-Cre+ mice; and Rictor?NKRaptorfl/+ represents Rictorfl/fl/Rptorfl/+/Ncr1-Cre+ mice. The dashed line indicates the control group, and the solid line indicates the gene knockout group. The MFI was calculated relative to total NK cells or to the CD27SP subset of NK cells from control littermates. The subpopulations of NK cells are distinguished by surface expression of CD27 and CD11b, and CD27SP, DP, and CD11bSP represent the CD27+CD11b-, CD27+CD11b+, and CD27-CD11b+ NK cell subsets, respectively. Each dot represents one mouse, and all experiments were replicated 3 (NK subsets in a), 4 (c, e–g), 5 (b), or 7 (total NK cells in a) times; error bars represent SD; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; one-way ANOVA (a–c, e–g). Control, n = 3; Rictor?/?, n = 5 in 3 independent experiments (d)

Fig 3: NK cell specification, maturation, and effector functions require mTORC1 activity. a Cumulative ratio and enumeration of NK cells (CD3-CD19-NK1.1+) present in BM, spleen, and peripheral lymph nodes (pLNs) from control (Rptorfl/+/Ncr1-Cre+) versus Raptor?NK (Rptorfl/fl/Ncr1-Cre+) mice. b Flow cytometric analysis and cumulative frequencies of NK cell (CD3-CD19-NK1.1+NKp46+) subsets found in the BM, spleen, and pLNs from control versus Raptor?NK mice. c Statistical quantification of IFN-?+ splenic NK cells (CD3-CD19-NK1.1+NKp46+) and the MFI of IFN-? among IFN-?+ NK cells following stimulation with PMA and ionomycin in the presence of GolgiPlug for 6 h. d Flow cytometric analysis and cumulative frequencies of CD107a+ cells among total splenic NK cells (CD3-CD19-NK1.1+NKp46+) and the indicated subpopulations thereof from control versus Raptor?NK mice after coculture with Yac-1 targets in the presence of GolgiStop for 6 h. e The cumulative ratio of RMA-S cell rejection 18 h after intraperitoneal injection of a 1:1 mixture of RMA and RMA-S cells into each mouse. f, g Flow cytometric analysis and cumulative data depicting the expression of perforin and granzyme B (f), CD98 and CD71, and uptake of 2-NBDG (g) within total splenic NK cells (CD3-CD19-NK1.1+NKp46+) and the indicated subpopulations thereof. The MFI was calculated relative to total NK cells or to the CD27SP subset of NK cells from control littermates. CD27SP, DP, and CD11bSP represent CD27+CD11b-, CD27+CD11b+, and CD27?CD11b+ NK cell subsets, respectively. Each dot represents one mouse, and the experiments were replicated 3 (c), 4 (g), 5 (a, b), 7 (d), or 9 (f) times; error bars represent SD; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; unpaired two-tailed Student’s t test with Welsh’s correction (a–d, f, g); generalized linear models (e). Each of the shapes represents a matched pair of littermates in an individual experiment, and the horizontal red line represents the median (e)

Fig 4: mTORC2 negatively regulates NK cell cytotoxicity against tumor cells. a Flow cytometric analysis and cumulative frequencies of CD107a+ cells among total splenic NK cells (CD3−CD19−NK1.1+NKp46+) and within the indicated NK cell subpopulations from control versus Rictor∆/∆ (top) or control versus Rictor∆NK (bottom) mice after co-culture with Yac-1 targets in the presence of GolgiStop for 6 h. b The specific lysis of RMA-S tumor cells when co-cultured with NK cell subsets from control versus Rictor∆/∆ or control versus Rictor∆NK mice for 4 h at an E:T ratio of 1:1; NK cell subsets were sorted based on the expression of CD27 and CD11b, while RMA-S cells were labeled with Calcein-AM. c, d Flow cytometric analysis depicting perforin (c) and granzyme B (d) expression by splenic NK cells (CD3−CD19−NK1.1+NKp46+) and subpopulations thereof from control versus Rictor∆/∆ (top) or control versus Rictor∆NK mice (bottom) mice. e The cumulative ratio of RMA-S cell rejection 18 h after intraperitoneal injection of a 1:1 mixture of Cell Trace Far Red-labeled RMA and Cell Trace Violet-labeled RMA-S cells into control versus Rictor∆/∆ or control versus Rictor∆NK mice. f–h Flow cytometric analysis and cumulative results depicting expression of CD98 (f), CD71 (g), and the uptake of 2-NBDG (h) by total splenic NK cells or the indicated NK cell subsets, comparing control versus Rictor∆/∆ (top) or control versus Rictor∆NK (bottom) mice. The MFI was calculated relative to total NK cells or the CD27SP NK cell subset from control littermates. CD27SP, DP, and CD11bSP represent CD27+CD11b−, CD27+CD11b+, and CD27−CD11b+ NK cell subsets, respectively. For the bar graphs, each dot represents one mouse, and all experiments were replicated 5 (h bottom), 6 (a bottom, f, g, h top), 7 (c bottom, d bottom), or 8 (a top, c top, d top) times; error bars represent SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001; unpaired two-tailed Student’s t test with Welsh’s correction (a, c, d, f–h); generalized linear models (b, e). Each of the shapes represents a matched pair of littermates in an individual experiment, and the horizontal red line represents the median (b, e)

Fig 5: Mechanisms of mTORC2 repression of mTORC1 activity in NK cells. a Intracellular flow cytometric analysis depicting the phosphorylation (p-) of Akt at Thr308 (p-AktThr308) in total splenic NK cells (CD3−CD19−NK1.1+). b The expression of p-S6ser235/236 in splenic NK cells from control versus Rictor∆/∆ mice after treatment with DMSO (set as the solvent control) or 2 μM AKTi-1/2 for 6 h ex vivo. c The abundance of CD122 expression on splenic NK cells (CD3−CD19−NK1.1+NKp46+) from control versus Rictor∆/∆ mice, analyzed by flow cytometry. d Statistical quantification of p-STAT5Y694 in splenic NK cells (CD3−CD19―NK1.1+) with or without IL-15 stimulation for 4 h ex vivo. e–g, i Statistical quantification of p-S6ser235/236 (e), perforin and granzyme B (f), frequencies of CD107a+ NK cells (g), SLC7A5 expression (i) in splenic NK cells from control versus Rictor∆/∆ mice after treatment with DMSO or 100 μM STAT5-IN for 6 h ex vivo. h Statistical quantification of SLC7A5 expression on both total NK cells and the indicated subsets. j The expression of p-S6ser235/236 in splenic NK cells from control versus Rictor∆/∆ mice after treatment with DMSO or 25 mM BCH for 6 h ex vivo. The MFI was relative to the littermates of control + DMSO group (b, e, f, i, j). Each dot represents one mouse, and all experiments were replicated 3 (a, b, e–j), 5 (d), 6 (c) times; error bars represent SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; unpaired two-tailed Student’s t test (a, c, h); one-way ANOVA (b, d–g, i, j)